Hi Reid,
It looks like the project you are trying to import only contains one document. The clustering algorithm
is failing, because it is expecting a set of documents.
If the document consists of multiple documents in one file, you can try to manually break the PDF into
multiple documents (or break by page), upload them to DocumentCloud, and then import the new project.
In the future, we hope to provide functionality that automate this process.
Our next release (on Monday), contains a fix that will provide a more helpful error message in this case.
_jonas